Developing Of Modeling And Simulation Of The Multiple Distillation Column With Downcomer

Crude oil distillation is the leading and critical unit of the refinery and petrochemical business. It is a important factor to economic interest of enterprises whether the distillation column can run safe, stable and efficient. But the proposes can not allowed to be applied to the practical process, because distillation is a lager-scale process. The first step must be done on the computer simulation platform. In order to develop the simulation system, the theoretical research of distillation process has been done, and then the simulation was set up to meet the actual needs. Main tasks of the paper are summarized as follows:(1) Dynamic simulation model was established on the basis of in-depth analysis for distillation process principle. Most of the distillation column models ignore the impact of downcomer and the pressure change of each column plate. These assumptions simplify the calculation, but can also lead to too poor a match between simulation results and the real situation to satisfy the practical requirements. Aiming at the above issues, this paper made such correction as follows. They are, correcting the dynamic mechanism model according to the downcomer influence; integrating the tower pressure changes into the model according to the close relationship between the tower pressure and gas load.(2) An algorithm that using Least Squares Support Vector Machine to predict bubble points of multicomponent mixtures was proposed, aiming at the iterative calculation in phase equilibrium calculation. Gear integration algorithm of variable step and variable order is effective to solve the model stiff problems; but for the calculation of multicomponent VLE, iterative algorithm can not satisfy the request of real-time dynamic simulation, because of the huge computation, long computation time and so on. Therefore, an algorithm that using least squares support vector machine to predict bubble points of multicomponent mixtures was adapted to improve the computational efficiency of the mechanism model.(3) Steady-state and dynamic simulations were carried out and a series of transition curves and temperature curves of distillation column component concentration change. The result shows that, the model can accurately simulate the actual distillation production process. And it provides high practical value and theoretical significance for control system analysis and simulation system design of the distillation.